难道见到这么有良心的帖子 - 一个meta eng director对perf review的反思

Admire这么有勇气的说出这么一番话哪怕自己也是既得利益者
哈哈，也顺便求个大米！https://www.linkedin.com/posts/r ... m_medium=member_ios

Fresh out of performance reviews at my company, I'm convinced there's a loophole in how performances are attributed to low ranked software engineers/ICs versus senior leaders and managers, that can potentially put ICs at a disadvantage.

Attributing success:
- Senior leaders can easily take credit for work done by their teams without a clear causality between their direct contributions and the team's achievements.
- ICs are typically expected to have direct evidence of their contributions that led to success, e.g. code written and reviewed, design documents and PRDs that led to successful project/products, etc.

Attributing failure (or the euphemism 'area of development'):
- Leaders and managers are rarely held accountable for poor decisions made along the way, especially if the overall project or product is broadly seen as successful.
- ICs are easily held accountable for causing SEVs or not generating direct impact, for not having written or reviewed enough code or design documents, etc.

In other words, there's a stronger causal relationship between an IC's documented work and their success, while for senior leaders and managers, it's easier to ride the wave of cherry-picked successes of their orgs. Conversely, it's relatively easy to attribute specific failures to ICs (SEVs, problematic code, lack of code/design, etc) compared to missteps and misdirections by a leader, as long as the overall area is successful.

IMHO the latter seem more problematic -- riding the wave of success might be fine at times, but not having a framework of accountability for leadership failures is worse. The time horizon for decision outcomes is often so long, attribution and accountability is tough. A timely example is over hiring and then doing layoffs -- the cycle is so long that it's virtually impossible to track down all the contributors to poor decisions (believe me, it's not just the CEOs).

No wonder so many ICs want to become managers even when that's not seen as a 'promotion' at their organization.

As someone with an evenly split career as IC and manager, I've seen this across companies. Running performance reviews for a few dozens to now my nearly 200 person org, I also see this at scale. This issue isn't talked about enough in the industry, let alone a fix attempted. Probably because we leaders are also often tasked with improving these systems. It's natural for defense mechanisms to kick in, leading to this being swept under the rug.

I think most medium-large companies suffer from this. I don't have a solution but did want to raise awareness. Our industry could take the lead in doing better, particularly in how we measure accountability of senior leaders. If we did, decisions would get made more carefully and thoughtfully, just like carefully constructed pieces of code prior to going into production.

补充内容 (2023-02-24 11:15 +8:00):
太长，问ChatGPT来个中文总结：

作者认为，在他们所在公司中，对于低级软件工程师/ IC和高级领导和经理的表现归因存在漏洞，这可能会对IC造成不利影响。作者认为高级领导很容易为团队的成功而取得功劳，而他们的直接贡献与团队的成就之间往往没有明显的因果关系，而IC通常需要有直接证据证明他们的贡献导致了成功，例如编写和审查的代码、设计文件和PRD等。作者指出，领导和经理很少因犯错而受到责任追究，特别是如果整个项目或产品被广泛认为是成功的，而IC很容易因造成严重漏洞或未产生直接影响、编写或审查的代码或设计文档不足等原因而受到追究。作者认为，IC的工作和成功之间存在更强的因果关系，而对于高级领导和经理来说，他们更容易沿着团队的成功道路前进。相反，将特定的失败归咎于IC相对容易，与领导的错误判断和错误方向相比，只要整个领域是成功的。作者认为后者更具问题性，因为在决策结果的时间跨度通常很长，因此归因和责任追究很困难。作者认为，业界没有足够讨论这个问题，更不用说试图解决它。他们认为，领导人通常也被赋予改善这些系统的任务，导致防御机制出现，问题被掩盖。作者表示，他们曾经在IC和经理之间平均分配职业生涯，并且在管理几十人到现在的近200人组织的绩效评估中也看到了这一问题的规模。作者没有解决方案，但希望引起关注。作者认为，我们的行业可以在如何衡量高级领导责任方面发挥带头作用。如果我们这样做，决策将变得更加小心谨慎，就像在投入生产之前精心制作的代码一样。

我之前的mgr升职的文档我看了，简直令人发指，就是单纯地罗列了组里senior eng在做的一些事，比如一个我lead的project，他很少参与，却把它当做他的credit

这个问题的本质还是performance review是上级给下级做的，而不是反过来。下级给上级做的叫做“Satisfaction survey”。有power在手的人会不自觉地利用自己的power和话语权维护自己的power，就像高级IC也会甩锅低级IC一样。

人可以背叛自己的阶级，但阶级不会背叛自己的阶级。一个有良心的个体反思也改变不了大局势。it行业下行的时代肯定会有这种怪现状，就像boom的时候很多公司无脑招人一样。
不是幸灾乐祸。我自己也是从业人员。我也希望有更多有良知的人可以发声。

大多数公司机构都这样

这是在为pip manager开路

不过level高的人如果决策错误，也比IC更容易被赶走

这是逼着大家以后都走上管理岗位，反正ic最后都是背锅侠。老板们抢credit甩锅都是高手

怎么感觉是在为即将到来的砍中层manager做好吹风工作

一个产品， 不是IC决定要做，设计一堆问题， 回头产品做不好了，不做了，不投资了，把那个IC裁了，这个合理吗？
还被披上low perf, not efficiency  骂名